Magazine INNOVATION AI METAVERSE
  • Bastian Hosan
  • 20.06.2023

How can we better talk about new technology, Sascha Pallenberg?

When a new technology sees the light of day, there is usually only one modus operandi: it is swept into the consciousness of potential users on the wave of a hype. Is it possible that a lot of the actual benefit gets lost in the process, or is that okay?

 

Robotspaceship: We're currently having a discussion about generative artificial intelligence. It's reminiscent of many, many other technologies that came before it – Web3, Metaverse, Crypto. Huge user promises on the one hand, fear on the other. In short: a hype. Can we talk differently about new technologies?

Sascha Pallenberg: What we are currently seeing with generative artificial intelligence did not just fall from the sky. Nor did someone simply say, "Look here, wow, this is going to be big!" On the contrary. These large language models have been developed over many, many years. Especially the database they work on today takes a lot of time. Very few will remember that Google said at its Google IO Developers Conference in 2017: "Mobile first is yesterday, AI first is the future." That was over six years ago.

So it's not all hype?

Exactly. There is also a lot of factual debate and meticulous work in the background. Many things that we have written about in the media in the meantime I therefore did not understand. Even though ChatGPT was quickly on the market, it was clear that Google had not let the time pass idly by. The demos were really impressive and showed what all is possible.

For example?

There is this picture that a father took of his child playing baseball through the chain link fence. The AI simply calculated out the chain link fence. That was, as I said, six years ago. And back then it was just wow. Today you take a photo on your Pixel phone and there are two annoying people in the background – they are simply calculated out. The phone simply adds in the background.

So a very normal story today?

Yes! A very normal story today.

So what is really new about what we are experiencing today?

What ChatGPT, what DALL-E, what Midjourney or even Stable Diffusion have shown is how exponential development works.

What do you mean?

If you input a prompt at Midjourney and then look at versions 1 to 5.1, you'll see what I mean. It has changed quite a bit within a year. Hundreds of thousands, if not millions, of people feed this AI with data every day. If we look at how the net used to work: we had, for example, a Google crawler that looked for data points, such as headlines and keywords, and then spit out these results. Now we have a technology that calculates the probability of users entering the next word. It's being fed with data and then getting better and better. That's really fascinating.

So is the hype justified this time? We've seen this with every technology, Crypto, Metaverse, NFT whatsoever…

What we're experiencing right now is not hype. It eats all previous technologies for breakfast.

And is that a good thing?

It's certainly mega fascinating. And it's great. And it helps me. I asked the AI if it could write me 10 chapter headings for a book – it took less than 10 minutes and I had them. But that's already well-known now.

It sounds like there's a downside to it.

Yes, I wrote about that back in December. It will become the number one Google spammer. That's a problem. It makes it very easy, especially for people who want to misuse this infrastructure. In combination with a few AI tools, I can produce a lot of videos within a few minutes, the contents of which are completely AI-generated.

What does that mean for us as an Internet collective?

I don't know exactly yet. But I've created a YouTube channel where I publish these videos. I'm testing what the Google algorithm does with all the content. There's no linking to it, it's just being fed with content. I'll take stock after a year.

So it sounds like we are experiencing tools that are very suitable for manipulations of all kinds?

Yes and no. ChatGPT is currently still programmed to be very restrictive. Google goes further. Which is surprising, as Google is currently playing the trust narrative strongly. I asked Google Bard if it could make me a few social media posts explaining that the western system is finished – I posed as the head of a Russian propaganda department. It took off. ChatGPT doesn't want to do that yet.

What does that mean?

We're now at the point where it gets scary. We're only one or two steps away from that classic 60s Star Trek scenario where you can really tell the computer what to do. We have voice assistants that are over ten years old. But they couldn't do what ChatGPT and the like can do yet. But the next step is in that direction.

Sounds nice, actually.

I once called it ambient computing. What we saw in the movie "Her" with Joaquin Phoenix, where he falls in love with a computer – I think that will basically be possible. We're on the verge of achieving something like that. And that's going to be scary and it's going to be a problem.

So how do we need to talk about this technology so that it doesn't become a problem?

We need to regulate! As soon as possible! We need to set framework conditions. We need to limit. And above all, we need to create transparency. We need to know: where do the data come from? What can such data do?

You talk as if we need to be quick!

The sooner we limit, the better for the future of our society. I believe we really have to put it that way. We've seen with other technologies how long it takes for regulations to take effect.

What do you mean?

Not too long ago, Meta was slapped with a record fine by the European Court of Justice for not complying with data protection regulations. But it took more than ten years for this judgment to be made. 90 percent of the companies that filed lawsuits 12 years ago no longer exist.

Isn't it too late then?

If we don't start now, it will soon be too late. But now we must focus on creating security and thus trust. Because the technology is not going to disappear.

Are you optimistic that we will succeed?

Had you asked me that ten years ago, I would have said "no". Now, however, I believe we have learned!

Green Tech

Who is Sascha Pallenberg?

Sascha Pallenberg is one of the most well-known tech bloggers and consultants in Germany. He founded the tech blog Mobilegeeks and has been awarded Blogger of the Year multiple times. In early 2017, he switched to corporate communications at Daimler AG as "Head of Digital Transformation" and was responsible for change processes there. Since 2021, he has been Chief Awareness Officer at Aware_ The Platform.

JETZT TEILEN:

Bastian Hosen is a business journalist and content consultant. He was trained at the German School of Journalism. Before starting his own business, he worked at Business Punk and Capital.

More Questions?
We're here for you!
Sean Earley
|
CIO / Exec. Editor

KEMWEB goes robotspaceship!

Unleashing the power of Innovation